Kouki Mohamed Amine

Interests

Things I Love

AI Engineering

My journey into Artificial Intelligence evolved from curiosity to professional expertise. Working on cutting-edge projects like LLM-powered multi-agent systems with LangGraph, LangChain, and CrewAI, and Cache-Augmented Generation, I've experienced how AI transforms raw data into intelligent decisions. From reducing latency by 25% to building agent pipelines that execute structured tools, building systems that learn and optimize continues to fuel my passion.

Full-Stack Development

Full-Stack Development provides the foundation for bringing AI solutions to life. Whether crafting intuitive Angular interfaces or architecting robust Spring Boot backends, I bridge the gap between intelligent algorithms and user-friendly applications. My experience spans from medical management platforms to internship systems, always focusing on seamless integration and optimal performance.

Machine Learning & Computer Vision

Diving deep into CNNs, Transformers, and deep learning frameworks like PyTorch and TensorFlow has opened endless possibilities. From achieving 93% accuracy in image classification to implementing RAG systems for real-world applications, I'm passionate about training models that solve tangible problems and deliver measurable results.

Open Source & Innovation

Contributing to the tech community through open-source projects and competitions (like achieving 2nd place at Pitch for Change 5.0) keeps me connected to innovation. I believe in collaborative development, sharing knowledge through platforms like Kaggle and HuggingFace, and building technology that advances the entire field while remaining accessible to all.

But who am I?

About Me

Hey there, I'm Mohamed Amine Kouki, an AI Engineer and Computer Engineering graduate from the National Engineering School of Carthage (ENICarthage). My professional journey spans AI optimization, full-stack development, and intelligent system design.

Currently, I'm an AI Engineer at Originator, working on LLM-powered agent systems. Before that, I built multi-agent automation pipelines with LangGraph, LangChain, CrewAI, and MCP at Hexagone SAS, and earlier optimized an AI platform there using Cache-Augmented Generation — improving memory utilization by 40%, cutting response time by 25%, and reducing costs by 30%.

My expertise spans LLMs, Transformers, and computer vision, alongside production-ready application development with modern frameworks like Angular, Spring Boot, Vue.js, and Flask — from secure medical management platforms to internal tooling, I focus on creating solutions that make real impact.

Beyond professional work, I'm active in the tech community—competing in innovation challenges, contributing to open-source projects, and continuously exploring emerging technologies. When I'm not coding, you'll find me playing chess or diving into robotics. Let's connect and create innovative AI-powered solutions together!

Where I've worked

Experience

AI Engineer

Jul 2026 – Present

Originator

Building and deploying LLM-powered agent systems.

AI Engineer

Oct 2025 – Mar 2026

Hexagone SAS · Tunisia

Developed and deployed scalable AI systems for document automation and enterprise workflows using modular Python backends and intelligent AI agents.

Built multi-agent workflows using LangGraph, LangChain, CrewAI, and MCP
Designed and optimized RAG pipelines for intelligent document processing
Integrated MCP tools into agent pipelines for structured tool execution

Python · LangGraph · LangChain · CrewAI · MCP · RAG · Docker

Freelance AI Developer

Jul 2025 – Oct 2025

Remote · Tunisia

Developed an AI-powered nutrition recommendation system using Retrieval-Augmented Generation (RAG) to provide personalized meal planning and dietary guidance based on individual goals.

Built a system calculating macros and tracking meals
Integrated an LLM for generating optimized meal suggestions
Delivered an end-to-end solution from data cleaning to deployment

Python · RAG · LLM · Vector Database · Pandas

AI Engineering Intern · Final Year Project

Feb 2025 – Jul 2025

Hexagone SAS · Tunisia

Improved an AI platform for scientific claim extraction via Cache-Augmented Generation (CAG), reducing latency and improving accuracy over traditional RAG.

Improved memory utilization by 40% through advanced caching
Reduced response time by 25% while boosting answer accuracy
Deployed the optimized system in production, cutting costs by 30%

Python · PyTorch · CAG · RAG · BERTScore · Docker · Hugging Face

What do I know?

Skills

Here are some of the technologies that are part of my current stack:

AI & Machine Learning

Python

PyTorch

TensorFlow

Keras

NumPy

Pandas

LangChain

LangGraph

CrewAI

MCP

Frontend Development

Angular

Vue.js

HTML

CSS

JavaScript

TypeScript

Bootstrap

Backend Development

Spring Boot

Java

JavaFX

C & C++

Databases

MySQL

PostgreSQL

Firebase

PL/SQL

What I uses?

Tools

Here are some of the tools that are part of my current stack:

VSCode

IntelliJ

PyCharm

Colab

Kaggle

HuggingFace

Docker

jenkins

Github

Postman

Figma

Canva

VirtualBox

Arduino

What I've done?

Projects

RAG Nutrition System

Python| RAG| LLM| Data Cleaning

Kaggle

Freelance : A personalized AI-powered meal planning system that calculates user macros, tracks consumed meals, and generates optimized meal suggestions based on goals, preferences, and calorie targets.

Optimizing AI with CAG

Python| PyTorch| CAG| RAG| LLM| Docker

Final Year Project : Optimized AI platform using Cache-Augmented Generation for scientific claim extraction. Achieved 40% better memory efficiency, 25% faster responses, and 30% cost reduction.

Date Image Classification

Python| TensorFlow| Keras| CNN| Kaggle

Kaggle

Developed a Convolutional Neural Network to classify date images into specific categories, achieving 93% accuracy. Utilized data augmentation, architecture optimization, and hyperparameter tuning to improve model performance.

Internship Management

Spring Boot | Angular | Bootstrap| MySQL

Github

A Multi-User web application for managing final year project topics at a university.

Medical Center

Vue.js | Firebase | Firebase-hosting

Github

Multi-user web app for communication between patients, doctors, and admins. Access the app here

Sports Complex

JavaFX | CSS | MySQL | FXML

Github

An offline desktop application with an interactive dashboard and a calendar dedicated to each field.

Cafe Management

Spring Boot | Angular | Bootstrap| MySQL

Github

A Multi-User Web Application for Managing a Cafe-Restaurant Shop

Vue.js Application

Vue.js | Firebase

Github

This project is a simple Vue.js application integrated with a Firebase backend. It supports multiple users, allowing administrators to manage user statuses .

Fruit Detection

CNN | TensorFlow|Streamlit

Github

This project leverages a Convolutional Neural Network (CNN) model to detect and classify fruit images into multiple categories .

Credentials

Certifications

Building Reliable Agents

LangChain Academy · Mar 2026

View Certificate

Introduction to LangChain

LangChain Academy · Mar 2026

View Certificate

Advanced Learning Algorithms

DeepLearning.AI & Stanford Online · Nov 2024

Verify

Supervised Machine Learning: Regression and Classification

DeepLearning.AI & Stanford Online · Nov 2024

Verify

Kouki Mohamed Amine

Interests

Things I Love

But who am I?

About Me

My CV

Resume

Where I've worked

Experience

AI Engineer

AI Engineer

Freelance AI Developer

AI Engineering Intern · Final Year Project

What do I know?

Skills

AI & Machine Learning

Frontend Development

Backend Development

Databases

What I uses?

Tools

What I've done?

Projects

Credentials

Certifications

Building Reliable Agents

Introduction to LangChain

Advanced Learning Algorithms

Supervised Machine Learning: Regression and Classification

What's next!

Get In Touch